1 vista deep extragalactic observations (video) surveyvideo/public/elais-s1_files/... ·...

ESO Public Surveys Proposal Form Phase 1

1 VISTA Deep Extragalactic Observations (VIDEO) Survey

Matt J. Jarvis, Univ.Oxford / Univ. Herts, UKCASU (VDFS; Cam), WFAU(VDFS; Edin), Omar Almaini(Notts), Carlton Baugh(Durham),Eric Bell(MPIA), Malcolm Bremer(Bristol), Jarle Brinchmann(Porto), Karina Caputi(IAS),Andrea Cimatti(Arcetri), Michele Cirasuolo(Edin), Lee Clewley(Oxf), Roger Clowes(UCLan),Chris Collins(LivJM), Garret Cotter(Oxf), Scott Croom(AAO), Gavin Dalton(Oxf), JamesDunlop(Edin), Alastair Edge(Durham), Jim Emerson(QMUL), Eduardo Gonzalez-Solares(Cam),Natascha Foerster-Schreiber(MPE), Marijn Franx(Leiden), Olivier Le Fevre(OAMP), SimonHodgkin(Cam), Rob Ivison(Edin), Mike Jones(Oxf), Scott Kay(Oxf), Jean-Paul Kneib(OAMP),Matt Lehnert(MPE), Jon Loveday(Sussex), Dieter Lutz(MPE), Bob Mann(Edin), RossMcLure(Edin), Klaus Meisenheimer(MPIA), Bob Nichol(Ports), Ian McHardy(Soton), RichardMcMahon(Cam), Mariano Moles(IAA), Seb Oliver(Sussex), John Peacock(Edin), Will Per-cival(Ports), Steve Rawlings(Oxf), Tony Readhead(Caltech), Hans-Walter Rix(MPIA), KathyRomer(Sussex), Piero Rosati(ESO), Huub Rottgering(Leiden), Stephen Serjeant(OU), RobSharp(AAO), Chris Simpson(LivJM), Ian Smail(Durham), Will Sutherland(Cam), BramVenemans(Cam), Aprajita Verma(Oxf), Montse Villar-Martin(IAA), Ian Waddington(Sussex),Fabian Walter(MPIA), Nic Walton(Cam), Chris Wolf(Oxf)

1.1 Abstract:(10 lines max)

The VISTA Deep Extragalactic Observations (VIDEO) survey is a 15 sq. degree, Z,Y,J,H,Ks survey specificallydesigned to enable galaxy and cluster/structure evolution to be traced as a function of both epoch and envi-ronment from the present day out to z=4, and AGN and the most massive galaxies up to and into the epoch ofreionization. With its depth and area, VIDEO will be able to fully probe the epoch of activity in the Universe,where AGN and starburst activity were at their peak and the first galaxy clusters were beginning to virialise.VIDEO therefore offers a unique data set with which to investigate the interplay between AGN, starbursts andenvironment, and the role of feedback at a time when it is most crucial. The multi-band nature of the surveyensures many key science drivers can be tackled using the survey alone, without recourse to data from otherwavebands. However, the survey fields have been carefully selected to ensure a good RA spread and mix of fieldswith existing multi-band data thereby enhancing the usefulness of the survey to the whole of the astronomicalcommunity, and with an eye to future use of other ESO facilities such as APEX and ALMA. The area anddepth means that VIDEO fits naturally between the proposed VIKING and Ultra-VISTA surveys, maximisingthe legacy to the ESO community and worldwide.

2 Description of the survey: (Text: 3 pages, Figures: 2 pages)

2.1 Scientific rationale: The Aims of VIDEO

We are already at a point where we have excellent constraints on the spatial distribution and properties ofgalaxies in the local Universe from three of the largest surveys ever undertaken, namely the 2dF galaxy redshiftsurvey (2dFGRS; Colless et al. 2001) and the Sloan Digital Sky Survey (SDSS; Adelman-McCarthy et al. 2006)in the optical and the IRAS redshift survey (Saunder et al. 2000) selected in the far-infrared. With the VISTAHemisphere survey this will be pushed out to z ∼ 0.6, and the VISTA-VIKING survey will reach z ∼ 1 for anL? galaxy. The aim of the VIDEO survey is to gain a commensurate data set at 1 < z < 4 overa similar volume to the surveys covering lower redshifts. This will allow galaxy evolution to betraced over the majority of the Universe, and from the richest clusters to the field.

The VIDEO survey will be driven by the VISTA data alone, without the pre-requisite of additional data inother wavebands. However, the fields are chosen to incorporate current and future multi-wavelength data setsto facilitate the broadest exploitation of the VIDEO survey data both within ESO and across the globe. Themain scientific aims of VIDEO are described below.

ESO-VISAS ([email protected]) page 1 of 15


2.1.1 Tracing the evolution of galaxies in all environments over 90% of the Universe

How and when were massive galaxies formed? When did they assemble the bulk of their stellar mass and how?Where does this mass assembly occur? These are crucial questions to which we still need answers.

(i) Galaxy formation over the epoch of activity

The advent of deep multi-wavelength surveys has led to a huge progression in this field. There are two waysto measure the build-up of stellar mass: by directly observing star formation, or by measuring the total stellarmass already formed by a given epoch. The first technique was revolutionised by the Lyman break technique(Steidel et al. 1996) which found large numbers of star forming galaxies at z ∼ 3. More recently GALEXhas stretched this to lower redshifts (e.g. Martin et al. 2005). However, these techniques are only sensitive torelatively unobscured systems, and are known to miss much of the luminosity density arising from galaxies atearly epochs.

On the other hand, mass estimates of distant galaxies require long-wavelength observations to probe rest-framewavelengths which are not dominated by young stars with low mass-to-light ratios. Surveys with Spitzer, and inparticular the 49 sq.deg. covered by SWIRE (Lonsdale et al. 2004) have allowed substantial recent progress inthis area (e.g. Babbedge et al. 2006). However, not even Spitzer has the combination of being able to reach bothto deep limits and wide areas as it is a small telescope with a relatively small field-of-view. This is borne out bythe fact that SWIRE only reaches z ∼ 2 for an L? galaxy. Surveys such as GOODS do probe to much greaterdepths but only cover very small sky areas. This has also been the problem with previous ground-based surveyslike FIRES (Franx et al. 2003). Up until now, these surveys have led the way in probing galaxy evolution fromthe earliest times up until the present day, but they are fundamentally limited by the fact that they cannotprobe scales larger than a few Mpc, severely limiting investigations of the environmental dependence of galaxyformation and evolution. As recently shown, cosmic variance can be significant for even moderately large surveyareas (e.g. the 0.8 deg2 of the CFHTLS; Ilbert et al. 2006). But VISTA is ushering in a new era of cosmologyby enabling us to rapidly survey representative volumes of the high-redshift Universe.

This is the goal of VIDEO, which will perform an in-depth study of the Universe during the crucial era 1 < z < 4,linking the shallower surveys (like VHS and VIKING) with the Ultra-VISTA survey. The depths of the VIDEOsurvey have been carefully chosen to reach L? at z = 4, and 0.1 L? at z = 1, thereby enabling us to detect thebulk of the luminosity density arising from galaxies over 90% of the Universe and the most massive galaxiesat the highest redshifts. Thus we will be able to investigate in exquisite detail which galaxies are in placefirst, and address the issue of downsizing in the mass function of forming galaxies where the massive early typegalaxies appear to be in place before the less massive galaxies (e.g. Cimatti, Daddi & Renzini 2006). It is alsoimportant to note that the epoch over which VIDEO is aimed is a crucial one in the history of the Universe, asthis is when the bulk of the star-formation and accretion activity took place (e.g. Madau et al. 1996; Steidelet al. 1999; Ueda et al. 2003; Richards et al. 2006) and so is the ideal survey with which to investigate theeffects that star-formation and accretion activity have on galaxy evolution in general. Moreover, the intrinsicrarity of the most luminous AGN, starburst and elliptical galaxies means that it is important to survey a largeenough area from which the luminosity function and clustering of particular objects can be constrained, thusenabling the evolution of bias to be determined. Crucially, VIDEO will have sufficient area to carry out theseinvestigations as a function of both redshift and environment. The area of VIDEO ensures that we are able toexplore ∼ 2 × 107 h−3 Mpc3 at z < 1, where we are sensitive to galaxies with L < 0.1 L?. Fig. 3 shows thatwith this survey we will detect ∼ 200 haloes with M > 1014 M at z < 1 (see also Fig. 4) and ∼ 2 haloeswith M > 1015 M, and as described in section 2.2 a significant sample of massive dark matter haloes at muchhigher redshifts. Thus, VIDEO will be able to explore the progenitors of all environments we see in the Universetoday.

Moreover, VIDEO will not only be able to detect the galaxies which contribute the bulk of the luminosity densityat these redshifts, but its 5 near-infrared filters will produce photometric redshifts accurate to ∆z/(1+z) ∼ 0.15(see Fig. 1), and to a higher degree of accuracy with ancillary optical data (see section 7). This will provide themost advanced data set for studies over the whole of the epoch of activity.

(ii) Statistical studies of the most massive galaxies at the highest redshifts



In addition to probing the sub-L? population up to z ∼ 4, VIDEO will be uniquely placed to quantify the densityand spatial clustering of the most massive galaxies at the earliest epochs. Recent work with the UKIDSS-UDSEarly Data Release show that the space density of z ∼ 5 galaxies with stellar mass M > 1011 M (McLure etal. 2006) is ∼ 18 per square degree (Fig. 2). This study was carried out with data ∼ 1 mag brighter than theproposed limit of VIDEO and over 30-times less area. Using the results of this study, means that we expect∼ 270 M > 1011 M galaxies at z ∼ 5 due to the increase in area, and extrapolating in redshift according toPress-Schechter (Bond et al. 1991) we expect ∼ 140 M > 1011 M galaxies at z ∼ 6 due to the increase inboth area and depth (Fig. 2). Thus we will be able carry out the first statistically significant clustering analysistowards the epoch of reionisation, providing a direct link between the underlying dark matter distribution andgalaxy populations and how this evolves up to the highest redshifts. We note that the Ultra-VISTA survey willdetect ∼ 10 times fewer M > 1011M galaxies at this important epoch, as the limiting factor is area rather thandepth, thus making clustering analyses impossible. However, we emphasise that the combination of VIDEO andUltra-VISTA will provide the premier data set for investigating the clustering properties of the highest redshift(z > 4) galaxies as a function of mass.

2.1.1 Tracing the evolution of clusters from the formation epoch until the present day

VIDEO also provides data over the area and depth with which to study the evolution of galaxy clusters fromtheir formation epoch to the present day. Galaxy clusters are essential tracers of cosmic evolution in the universefor two important reasons. First, clusters are the largest virialized objects whose masses we can measure. Massmeasurements of local clusters can determine the amount of structure in the Universe on scales around 1014 M.Consequently, comparisons of the present-day cluster mass distribution with the distributions at earlier epochscan be used to determine the rate of structure formation, placing constraints on cosmological models (see Fig. 4).Second, the deep potential wells of clusters also mean that they act as closed astrophysical laboratories thatretain their gaseous matter. Therefore clusters possess a wealth of information about the processes associatedwith galaxy formation such as the efficiency of which baryons are converted into stars and the effects whichfeedback processes have on galaxy formation.

At 0 < z < 1, clusters of galaxies appear to undergo little in the way of strong evolution, either in their gasphase properties or in the properties of the more massive galaxies within them (e.g. Rosati et al. 2002). Theepoch at z > 1 is therefore a crucial one for their evolution, which must be dramatic in the 4 Gyr between1 < z < 4. The design of VIDEO is such that it will be a crucial resource for the study of early cluster evolution.Its depth is such that it can trace the bright end of the cluster luminosity function to z = 3 and look in detail atthe less luminous cluster galaxies to z = 2, while its area should provide a sample of several tens of clusters withmasses above 1014 M at z > 1. Our recent study over ∼ 0.5 sq.degree (van Breukelen et al. 2006; Fig. 5) in theSXDF with UKIDSS-Ultra-deep survey and Spitzer-SWIRE data has shown that there are ∼ 5 M > 1014 Mclusters per square degree at z > 1, consistent with ΛCDM cosmologies (Fig. 4). However, this study is severelylimited by sample variance. Therefore larger areas are required in order to pin down the evolution of galaxieswithin clusters and the evolution in the cluster mass density at z > 1. With VIDEO we expect to discover∼ 75 massive clusters at z > 1 with the bulk of these (∼ 80%) at 1 < z < 1.5 (Fig. 4). Moreover, the depthof VIDEO coupled with the 15 sq.deg area ensures that we will also have a statistically meaningful sample ofmassive galaxy clusters at z > 1.5. In ΛCDM cosmologies we expect ∼ 1 M > 1014 M clusters per sq.deg atz > 1.5, thus in VIDEO we expect ∼ 15 massive clusters at a time when they are just beginning to virialise.

Several of the fields chosen for VIDEO have (or will have) complementary multi-wavelength data that will allowthe infrared properties of the cluster galaxy population (from VIDEO and SWIRE) to be linked to the gasphase of the clusters, most notably in the XMM-LSS, ELAIS S1 and CDF-S fields whose excellent X-ray dataare deep enough to identify the most massive clusters to z > 1. The VIDEO survey areas will also be withinreach of the new Sunyaev–Zel‘dovich telescopes (e.g. AMI, APEX, Bolocam, CBI2, SPT and the SZ-array; seeCarlstrom et al. 2002 for a review) which will be able to detect any virialized clusters down to a mass limit of∼ 1014M, essentially independent of redshift, thus perfectly complementary to VIDEO. Using cluster surveysto trace the evolution of Dark Energy in the Universe has the potential to exceed the expectations of BaryonAcoustic Oscillation and Supernovae techniques, but they also have the largest systematic errors (see report ofthe Dark Energy Task Force; Albrecht et al. 2006). To overcome these systematic errors we need to understand



the astrophysics of cluster evolution, and thus VIDEO data will be crucial in combination with X-ray and SZsurveys, along with spectroscopic follow-up.

2.1.3 Accretion activity over the history of the Universe

(i) The first accreting black holes

The SDSS has revolutionised studies of QSOs at the highest redshifts, and it provided the first evidence thatthe epoch of reionization was coming to an end around z>∼6 (Becker et al. 2001). Pushing to higher redshiftsis impossible with optical surveys, regardless of depth, due to the fact that the Gunn-Peterson trough occupiesall optical bands at z > 6.5. Therefore, to push these studies further in redshift needs deep wide-field surveysin the near-infrared. VIDEO’s combination of depth and area provides the ideal way in which toplace the first constraints on the luminosity function of z > 6.5 QSOs.

Up until now only the brightest QSOs at z > 5 have been found using the SDSS. In the future the UKIDSSLarge Area Survey and wide-area VISTA public surveys such as VIKING and the VISTA hemisphere surveywill probe the bright end of the QSO luminosity function at z > 6. However, the shape of the QSO luminosityfunction at these redshifts can only be studied with much deeper near-infrared imaging over a significant surveyarea. This is the only direct way to determine the contribution of accreting black holes to the reionization ofthe Universe and constrain the density of black-holes within the first Gyr after the Big Bang, as just using thebrightest quasars requires a huge extrapolation to fainter magnitudes.

The VISTA Z-band samples rest-frame wavelengths blueward of Lyα for objects with z > 6.5, so we can identifythese objects using the standard dropout technique. However, cool dwarf stars are also very red in Z−Y but canbe differentiated from high-z QSOs by their Y −J colours (Fig. 6; Hewett et al. 2006). QSOs have Z −Y > 1.5and Y − J < 0.8, and so we can make the distinction for objects brighter than our Z = 25.2 limit if we reachY = 23.7 and J = 23. Extrapolating the QSO luminosity function of Fan et al. (2001) to Y = 23.7, we expectz > 6.5 QSOs to have a surface density of ∼ 1 deg−2 (Fig. 7), thus there will be ∼ 15 QSOs within the epochof reionization over the full VIDEO survey area.

(ii) The peak of accretion activity in the Universe

Where optical data exist, VIDEO can find lower-redshift QSOs which are faint and/or reddened. However,more interesting constraints could be made on the obscured AGN population when the VIDEO survey data iscombined with Spitzer-SWIRE, X-ray and Herschel observations. The depth of VIDEO is crucial in determiningthe nature of distant obscured AGN. Although the central quasar is obscured in these objects, VIDEO will havethe sensitivity to detect the host galaxies of such sources up to ∼ 5 assuming that their host galaxies are∼ 2 − 3 L?. However, these sources are rare. Given current work over the ∼ 4 sq.deg of the Spitzer FirstLook Survey (Martinez-Sansigre et al. 2005; 2006; Bonfield et al. in prep.) we expect ∼ 300 heavily obscuredquasars at z > 2 in VIDEO. Therefore, VIDEO could in essence probe the cosmic evolution of all AGN, bothobscured and unobscured. Such a statistical study is beyond the reach of Ultra-VISTA due to the small area(expect ∼ 30), and VIKING due to its relatively shallow depth, as well as the UKIDSS-DXS due to having only2 filters and ∼ 1 mag brighter sensitivity. These intrinsically rare, but important objects, will be prime targetsfor ALMA as their CO masses will be large and hence dynamical masses could be measured over the history ofthe Universe.

To summarise, VIDEO data, coupled with Spitzer-SWIRE, X-ray, and in the future, Herschel data, will goalong way to providing a full census of AGN activity over the epoch when these sources were accreting at theirmaximum level. With the large area covered such activity can also be probed as a function of environment, aproblem which can only be addressed with the depth and area of VIDEO.

2.1.4 The VIDEO supernovae project

The VIDEO survey also provides an opportunity for a large-volume supernova (SN) search in a random volume ofthe low-redshift universe. In particular, near-infrared observations can be used select SNe almost independentlyof dust obscuration in their host galaxies. It will thus facilitate the best rate assessments for different SN typesand their dependence on host galaxy properties (e.g. mass, star-formation rate, metallicity). The VIDEO-SNe team led by C. Wolf and including several members of the former EU RTN on SN progenitors, including



Patat & Leibundgut at ESO, has developed a strategy for the SN search. With VIDEO, 250 core-collapseand 100 type-Ia SNe are expected during the 5-year project, at a median redshift of 0.2 in the ZY J-bands.Host galaxies will also be studied spectroscopically to constrain metallicities and unobscured star-formationrate. The multi-wavelength data, such as Spitzer and Herschel will also be used to assess star formation ratesand stellar masses. This will allow us to check the consistency of SN-rate and star-formation rates and assesspossible metallicity-dependencies in SN rates of all types. The inclusion of dust-obscured SNe in a full samplewill probably not increase the cosmic SN rates very much, but it will remove an important bias that is stronglycorrelated with host galaxy properties. Possibilities of VISTA data being pipeline-reduced on Paranal to searchfor transients are currently being explored with ESO.

2.2 Immediate objective:

The principal aim of VIDEO is to provide a complete near-infrared data set of unprecedented depth and areafrom which galaxy formation and evolution as a function of both epoch and environment can be explored. Oneof the main advantages of VIDEO is that much of the science can be carried out with the VIDEO data alone dueto the target redshifts and the 5-band filter approach. This truly will be a VISTA-led survey. Ancillary datasets from Spitzer, Herschel, along with optical, X-ray, radio and SZ surveys will broaden the scientific impactof VIDEO across the ESO community and the world. We emphasise that the surveys carried out at theseother wavelengths are uniquely matched to the area of VIDEO. The 1000 sq.deg. of VIKING will never havematching data in these wavebands. Furthermore, the Ultra-VISTA survey will be limited by sample variancefor the rarest and most massive structures, these are exactly the structures which will preferentially be detectedwith X-ray and SZ surveys. VIDEO is therefore crucial for these studies.

3 Are there ongoing or planned similar surveys? How will the pro-posed survey differ from those? (1 page max)

The only other survey which is currently underway or planned in the near future with similar aims is theUKIDSS-Deep Extragalactic Survey (DXS). The UKIDSS-DXS aims to survey 35 deg2 over four patches ofsky to depths of K ≈ 20.8 and J ≈ 23. Three of the four DXS fields cover SWIRE areas, and only one ofthese (XMM-LSS) is visible to ESO telescopes. The XMM-LSS field also contains the UKIDSS-Ultra DeepSurvey field, thus it is likely that VIDEO will complete the larger XMM-LSS area on a shorter timescale thanUKIDSS-DXS will, leaving the DXS to concentrate on the northern SWIRE fields.

VIDEO also has other crucial differences to UKIDSS-DXS:

(i) The DXS will only observe to K = 20.8, roughly a magnitude brighter than VIDEO. In real terms thismeans that the current DXS only probes a passively evolving L? galaxy to z ∼ 2, whereas VIDEO extendsthis up to z ∼ 4 and ensures complete coverage of the ‘active epoch’. Probing a magnitude deeper also ensuresthat clusters of galaxies can be investigated in detail, both with the VIDEO data alone and also with follow-upobservations. The usual method of detecting cluster galaxies from a single colour requires the survey to probeto M? +2 at any given redshift. For DXS this means galaxy clusters can be probed in detail to z ∼ 0.8, whereaswith VIDEO this becomes z ∼ 1.5, where we expect the most massive clusters to begin to virialise.

(ii) UKIDSS-DXS only uses the J and K-bands over the full 35 square degrees. Therefore, with the DXSwe are only able to use the J − K colour in order to decouple which galaxies are at z > 1. With the fivenear-infrared filters of VIDEO, much tighter photometric redshift constraints can be made, for instance usinga simple template fitting approach, accuracies of ∆z/(1 + z) < 0.2 (see Figs 1 & 2) can be obtained and addingoptical and SWIRE data can reduce this to ∆z/(1 + z) ∼ 0.1 over 0 < z < 5. Thus, we will no longer have torely on colour–magnitude relations to pick out candidate high-redshift clusters, as we will have a 3-dimensionalview. The Oxford team has already applied these techniques to the UKIDSS-UDS Early Data release (whichonly covers <∼1 deg2) with great success (van Breukelen et al. 2006). VIDEO is the only survey which has thedepth, area and filter combination to find the rarest and most massive overdensities in the Universe from z ∼ 4



Fig. 1a (left) Accuracy of photometric redshifts with VIDEO-ZYJHK alone. One can see that the photometricredshifts are well constrained over the full range in redshift that VIDEO aims to explore 1 < z < 4 even withoutoptical or Spitzer data with σ = 0.15. Fig. 1b (middle) Accuracy of photometric redshifts with VIDEO-ZYJHK +SWIRE-IRAC data. One can see that the photometric redshifts are well constrained over the full range in redshift(0 < z < 5) even without optical data, with a σ = 0.25. Fig. 1c (right) The photometric redshift distribution forVIDEO-ZYJHK + SWIRE-IRAC + ugri optical imaging with AB magnitude=25 in each optical band. One cansee that the addition of this relatively shallow optical data set reduces the uncertainties at z < 1 dramatically(σ = 0.11 for 0 < z < 5). We note that deeper data than this already exists over CDF-S and similar depth dataexists over XMM-LSS. We note that for the redshift range 1 < z < 4 the photometric redshift uncertainties usingall bands are consistent with the photometric redshifts using the VIDEO data alone, i.e. σ ∼ 0.15.

Fig. 2 (left) The filled circle at z = 5.3 is the estimate for the number density of galaxies with stellar massesgreater than ' 1011 M from the study of McLure et al. (2006). The data-points at z ≤ 3 (open circles) arethe estimated number densities of galaxies with stellar masses M ≥ 1011 M from the GOODS CDFS K−bandselected sample of Caputi et al. (2006). The curves show the redshift evolution of the number density of darkmatter halos with masses M ≥ 1.5× 1012 M. The solid, dashed, dot-dash and dotted curves correspond to fourdifferent values of σ8 (0.90, 0.85, 0.80 & 0.75 respectively) Fig. 3 (right). Each curve indicates the variation withredshift of the comoving number density of dark matter haloes with masses exceeding a specific value M in thestandard ΛCDM model with ΩM = 0.3, ΩΛ = 0.7, h = 0.7 and σ8 = 0.9. The label on each curve indicates thecorresponding value of log(M/M) [Taken from Mo & White (2002)].



Fig. 4 (left) The space density of clusters per unit redshift interval as a function of redshift, for the τCDM(ΩM = 1) and ΛCDM (ΩM = 0.3, ΩΛ = 0.7) cosmologies. Clusters are defined as dark matter haloes withM200 > 1014 M/h. (Taken from the Virgo Consortium Hubble Volume simulations (Evrard et al. 2002). Fig. 5

(right) A multi-colour image of a z ∼ 0.8 galaxy cluster from the UKIDSS-UDS early data release combined withdeep optical data from Subaru (van Breukelen et al. 2006). We note that VIDEO will be able to take this sortof study out to z ∼ 2.

Fig. 6 (left) Z-Y vs Y-J diagram illustrating colours of simulated stars, elliptical galaxies and quasars. Thesimulated objects colour coded as follows: BPGS O-K dwarfs (blue) filled circles; M dwarfs (green) open circles;L dwarfs (orange) filled triangles; T dwarfs (maroon) open triangles; Burrows model cool brown dwarfs (purple)filled squares; elliptical galaxies 0.0 < z < 1.5, ∆z = 0.1 red track with inverse triangles; quasars 5.0 < z < 6.7,∆z = 0.1 black track with black circles with redshifts z = 6.5 and z = 6.6 marked. Quasar with redshiftsz > 6.5 can be distinguished from cool stars by their red Z-Y colour and blue Y-J colour. The region incolour-colour space occupied by high redshift quasars is outlined by the dashed lines. Fig. 7 (right) Number ofquasars at 6.5 < z < 7.3 expected per square degree within VIDEO with Y < 24(Vega) for various luminosityfunction parameters extrapolated from Fan et al. (2003). (The current limit on β is ∼ 3.2 at z > 6.5 andM? ∼ −25 at z ∼ 2). The lines from top to bottom indicate the prediction using a luminosity function withM? = −26,−25,−24,−23 and −22. Therefore, we expect to be able to place the first measured constraints onthe luminosity function at these redshifts.

2



to z ∼ 1.

The VISTA-VIKING survey only has the depth to explore the Universe at z < 1 for an L? galaxy. This ispartly the reason why the optical data is crucial for KIDS as the characteristic galaxy spectral features stilllie in the optical waveband. These characteristic spectral features are redshifted into the near-infrared bandsat z > 1, and is what makes VIDEO the ideal project for studying this epoch. Moreover, by surveying areascovered by Spitzer SWIRE we will have 9-band photometry tuned to obtain photometric redshifts at z > 1,similarly to KIDS+VIKING at z < 1. Additional moderate depth optical data from the VST would increasephotometric redshift accuracy of the whole survey further by reducing the amount of low-redshift contaminantsto the z > 1 populations. We note that many of our fields already have adequate optical waveband coveragefrom other surveys (e.g. CFHT-Wide, CTIO and VVDS).

VISTA Ultra-deep surveys have the depth to probe the Universe up to the highest redshifts, and will provideexcellent data sets for the study of galaxy evolution since the epoch of reionization. However, what they gain indepth they lose in area, such that probing evolution as a function of environment with these surveys is severelyhampered by cosmic variance (see Scientific Rationale).

4 Observing strategy: (1 page max)

Our strategy for observing the VIDEO fields is as follows. We will ensure, where possible, each full 1.6 sq.degtile will be completed to the required VIDEO-specific depth in all five near-infrared colours once observationson a given tile have been started.

Tiles will be started from the centre of each field and working outwards to fill the whole survey field. This willensure that there is always a VIDEO field available throughout the year over the five years of the survey.

The break down by Period would be as follows (see section 5.1 for full details of time justification).

Period Time (h) Mean RA Moon seeing TransparencyP79(Apr’07-Sep’07) 57 18hr Dark < 0.8 THN,CLRP79(Apr’07-Sep’07) 48 18hr Grey < 0.8 THN,CLRP79(Apr’07-Sep’07) 28 18hr Bright < 0.8 THN,CLRP79(Apr’07-Sep’07) 23 18hr Bright < 0.6 THN,CLRP80(Oct’07-Mar’08) 57 06hr Dark < 0.8 THN,CLRP80(Oct’07-Mar’08) 48 06hr Grey < 0.8 THN,CLRP80(Oct’07-Mar’08) 28 06hr Bright < 0.8 THN,CLRP80(Oct’07-Mar’08) 23 06hr Bright < 0.6 THN,CLRP81(Apr’08-Sep’08) 57 18hr Dark < 0.8 THN,CLRP81(Apr’08-Sep’08) 48 18hr Grey < 0.8 THN,CLRP81(Apr’08-Sep’08) 28 18hr Bright < 0.8 THN,CLRP81(Apr’08-Sep’08) 23 18hr Bright < 0.6 THN,CLR

The survey would continue with this strategy over the 5 years of the survey, resulting in a total of 174 nights(assuming 9 hour nights).

5 Estimated observing time:

We base our exposure times on the typical elliptical galaxy colours at z > 2, where we are sensitive to thebulk of the luminosity density arising from galaxies. Therefore, for our limit of KS = 21.7, this corresponds toH = 22.7, J = 22.7, and Z = 25.2. We also include observations in Y to enable us to perform high-redshiftQSO – dwarf star separation. To distinguish these we require Z − Y > 1.5 and Y − J < 0.8, therefore we wishto probe to a depth of Y = 24.



Assuming 5σ for a point source in a 2 arcsec aperture, with 0.8 arcsec (0.6 arcsec in Ks) seeing we request thefollowing (per source, magnitudes are Vega). † is the proposed UKIDSS-DXS depth in H over 5 square degrees.

Filter Time (h) Time (h) Time (h) 5σ 2′′ ap.mag. UKIDSS Seeing Moon Transparency(per source) (per tile) (full survey) AB Vega Vega

(no overheads) (+overheads) (+overheads)Z 17.5 60.8 570 25.7 25.2 – 0.8 D THN,CLRY 6.7 23.2 218 24.6 24.0 – 0.8 G THN,CLRJ 8.0 27.9 261 24.5 23.7 22.3 0.8 G THN,CLRH 8.0 29.4 276 24.0 22.7 22† 0.8 B THN,CLRKs 6.7 23.8 224 23.5 21.7 20.8 0.6 B THN,CLR

The survey fields are described below (the first three are Spitzer-SWIRE fields), depending on schedulingconstraints the number of tiles on each field could be adjusted. However, our preference is that the SWIREfields accrue the most coverage.

Field RA-DEC Total Number Total Area Total Timeof tiles (sq.deg) nights

ELAIS-S1 0034-43 3 4.5 52XMM-LSS 0218-05 3 4.5 52

CDF-S 0332-27 2 3.0 34.5VIDEO-1 1400+05 2 3.0 34.5

The VIDEO-1 field is a new extragalactic field to fill the RA gap for extragalactic science with a view to ALMAand other upcoming instrumentation. It is also part of the VVDS shallow survey. If VIDEO-1 is selected as afield then we will actively pursue Spitzer data similar to SWIRE over that area. The number of nights shownassumes 9 hr nights.

5.1 Time justification: (1 page max)

Our aim is to be able to detect a galaxy at the break of the elliptical galaxy luminosity function over 90% ofthe Universe, which corresponds to 0 < z < 4. This ensures that we are sensitive to one of the most importantepochs in the Universe, where star-formation and accretion activity were at a maximum.

We use the K−band luminosity function from 2MASS (Kochanek et al. 2001) where M? = −24.3. Assuming apassively evolving stellar population with a high formation redshift, in agreement with extremely red objects inthe GOODS data set (e.g. Caputi et al. 2005), then we expect a z ∼ 4 elliptical galaxy to have a total magnitudeof K = 21.2. Under the assumption that 40% of a z > 2 elliptical galaxy’s light is lost if a 2 arcsec aperture isused (see e.g. Jarvis et al. 2001) then the measured magnitude within this aperture would be K ∼ 21.7, whichdefines our 5σ Ks-band limit.

For the Ks-band observations we use a tighter seeing constraint to enable accurate star-galaxy separation to becarried out over all of our fields in at least one filter, and the natural seeing is better in Ks than in the bluerfilters.

We use the VISTA exposure time calculator to calculate all of our time requests. For the Ks band we useDIT=10 s with NDIT=6, and we use 2 × 2 microstepping to allow better Nyquist sampling for this seeingconstraint, and a 5 point jitter pattern. To complete a a third of a tile (i.e. 2 pointings) (to ensure reasonabletime per OB) with this strategy requires 2866 s (including overheads), which gives a 5σ point source sensitivityof Ks = 20 mag. Therefore to reach the full depth with this strategy requires 38 OBs. To reach the fullsensitivity over one tile requires 30 OBs each of 2866 s, which equates to 24 hours per tile, and to cover the full15 sq.degrees requires 224 hours.



For the ZYJH bands we relax the seeing constraint to 0.8 arcsec as star-galaxy morphological separation willbe carried out with the Ks-band observations. Therefore, we no longer require microstepping and use only a5-point jitter pattern in each case.

For the H-band observations we again use a 5-point jitter pattern with DIT=10 s and NDIT=6. Therefore tocomplete a single 1.6 sq.deg tile requires 2207 s (including overheads). To reach the full tile depth of H = 22.7(5σ, 2 arcsec aperture) requires 48 OBs of this length (29.4 hours including overheads). For the full 15 sq.degreesthis equates to 276 hours in total.

For the J− we use 30 s DITs and NDIT=2, this equates to 2090 sec to complete one full 1.6 sq.deg tile downto a 5σ point source sensitivity of J = 21.5. Therefore to reach the full depth of J = 23.7 requires 48 OBs or27.9 hours per 1.6 sq.degree tile. For the full 15 sq.degree survey area this equates to 261 hours.

In Y we only aim to reach Y = 24 (Vega) to ensure that high-redshift quasar candidates can be distinguishedfrom dwarf stars (Fig. 6), whilst also providing extra photometric accuracy for Z-drop-out galaxies. This isrelatively inexpensive and provides a large scientific benefit for z < 2 galaxies as well as the high-redshift quasarsearch. Using DIT=30 s and NDIT=2, with a 5-point jitter pattern means that to completely cover one tilerequires 2090 s (including overheads) which reaches Y = 22 for a complete tile. Therefore to reach the fullsurvey depth on a 1.6 sq.degree tile requires 40 seperate OBs or 23.2 hours (including overheads). Thus to coverthe full 15 sq.degree requires 218 hours.

The Z−band observations need to be deeper than in the other bands as this is the filter we will use to probeshortward of the 4000A at z > 1. In Z−band we use DIT=45 s and NDIT=1 with a 5-point jitter pattern.Covering a full tile twice (Exposure Loops=2) with this strategy requires 3128 s (including overheads) per tile,reaching a 5σ depth of Z = 22.8. To reach the full survey sensitivity of Z = 25.2 requires 70 OBs, taking61 hours per tile. Therefore, the full 15 sq.degrees requires 570 hours.

Therefore, the full survey requires a total of ∼1550 hours.

We note that currently ESO assumes 9 hour long nights (10 hours in winter and 8 hours in summer), howeveran extended night for the near-infrared observations carried out with VISTA would ensure that all of thesurveys could be carried out at increased efficiency. For instance, Ks−band observations can extend well intoastronomical twilight.



6 Data management plan: (3 pages max)

6.1 Team members:

Name Function Affiliation CountryM. Jarvis PI & OB Preparation Oxford/Univ.Herts. UKA. Edge OB Preparation Durham UKR. McLure OB Preparation Edinburgh UKA. Verma OB Preparation Oxford UKCASU(VDFS) Pipeline processing Cambridge UKCASU(VDFS) Data Quality Control-I Cambridge UKJ. Emerson VISTA PI & VDFS Coordinator + OB prep QMUL UKWFAU(VDFS) Science Archive Edinburgh UKWFAU(VDFS) Data Quality Control-II Edinburgh UKN. Walton VO Standards Cambridge UK

VIDEO Specific TasksL. Clewley Data Quality Control-III Oxford UKI. Smail Data Quality Control-III Durham UKE. Bell, MPIA Postdoc Data Quality Control-III MPIA, Heid. DE. Gonzalez-Solares Frame Stack Cambridge UKOxford Postdoc Frame Stack Oxford UKM. Jarvis, L. Clewley Final Catalogue Production Oxford UKI. Smail, K. Coppin Final Catalogue Production Durham UKK. Meisenheimer, Postdoc Final Catalogue Production MPIA, Heid. DJ. Loveday Final Catalogue Production Sussex UK

Other data productsO. Le Fevre VIDEO-VIMOS strategy OAMP FH. Rottgering VIDEO-LOFAR strategy Leiden NLS. Rawlings VIDEO-GMRT cat. production Oxford UKR. Ivison VIDEO-VLA/ATCA cat. production Edinburgh UKS. Oliver, I . Waddington VIDEO-SWIRE cat. production Sussex UKS. Oliver VIDEO-Herschel strategy Sussex UKS. Croom & R. Sharp VIDEO-AAOmega cat. production AAT OtherM. Bremer & K. Romer VIDEO-XMM cat. production Bristol, Sussex UKG. Dalton VIDEO-FMOS cat. production Oxford UKF. Walter VIDEO-(sub)mm-wave strategy MPIA, Heid. DT. Readhead, M. Jones, K. Romer VIDEO-SZ strategy Caltech, Oxford, Sussex USA,UKM. Moles & M. Villar-Martin VIDEO-EMIR strategy IAA Spain

The CASU (VDFS) team consists of Irwin, Lewis, Hodgkin, Evans, Bunclark, Gonzales-Solares, Riello. TheWFAU (VDFS) team consists of Hambly, Bryant, Collins, Cross, Read, Sutorius, Williams.

6.2 Detailed responsibilities of the team:

The observation planning team will be led by the PI (Jarvis) and include members of the VIDEO team fromEdinburgh (McLure), Durham (Edge), MPIA-Heidelberg (Meisenheimer), Oxford (Verma) and Sussex (Loveday& Oliver) and VISTA PI Emerson (QMUL). They are responsible for generating the OBs using the Survey AreaDefinition Tool and P2PP and for revising these and monitoring survey progress using a local Data QualityControl database as necessary. Experience shows that the full scientific validation is only possible when peoplestart trying to do science with the data. Thus we will also have a number of people from Oxford, Durham,



Edinburgh and MPIA carrying out the first checks on the pipeline-reduced data.

6.3 Data reduction plan:

The data reduction will be carried out using the VISTA Data Flow System (VDFS; Emerson et al. 2004, Irwinet al. 2004, Hambly et al. 2004), operated by the VDFS team, and augmented by Jarvis, Gonzales-Solares,Clewley, Smail and Bell, especially for product definition and product Quality Control. This team alreadyhas experience with wide-field near-infrared data as Jarvis, Smail, Clewley, Gonzales-Solares are all members ofeither the UKIDSS-UDS or DXS working groups, and Gonzales-Solares performed the individual frame stackingfor the UKIDSS-DXS.

6.3.1 Pipeline processingThe Cambridge Astronomy Survey Unit (CASU) are responsible for the VDFS pipeline processing componentwhich has been designed for VISTA and scientifically verified by processing wide field mosaic imaging data fromUKIRT’s NIR mosaic camera WFCAM and is now routinely being used to process data from the WFCAM ata rate of up to 250GB/night. It has also been used to process ESO ISAAC data e.g. the FIRES survey dataand a range of CCD mosaic cameras.

The pipeline includes the following processing steps but is a modular design so that extra steps are easilyadded. All the steps will have been tested on a range of input VISTA datasets. and include: instrumentalsignature removal – bias, non-linearity, dark, flat, fringe, cross-talk; sky background tracking and removalduring image stacking – possible need to also remove other 2D background variations from imperfect multi-sector operation of detectors; define and produce a strategy for dealing with image persistence from precedingexposures; combine frames if part of an observed dither sequence or tile pattern; consistent internal photometriccalibration to put observations on an approximately uniform system; basic catalogue generation includingastrometric, photometric, shape and Data Quality Control (DQC) information; final astrometric calibrationfrom the catalogue with an appropriate and World Coordinate System (WCS) in all FITS headers; basicphotometric calibration from catalogue using suitable pre-selected standard areas covering entire field-of-viewto monitor and control systematics; each frame and catalogue supplied with provisional calibration informationand overall morphological classification embedded in FITS files; propagation of error arrays and use of confidencemaps; realistic errors on selected derived parameters; nightly extinction measurements in relevant passbands;pipeline software version control – version used recorded in FITS header; processing history including calibrationfiles recorded in FITS header

6.4 Expected data products from the VDFS:Instrumentally corrected frames (pawprints, tiles etc) along with header descriptors propagated from the in-strument and processing steps (science frames and calibration frames)Statistical confidence maps for each frameStacked data for dithered observationsDerived catalogues (source detections from science frames with standard isophotal parameters, model profilefitted parameters, image classification, etc.)Data Quality Control databaseDatabase-driven image products (stacks, mosaics, difference images, image cut-outs)Frame associations yielding a survey field system; seamless, merged, multi-colour, multi-epoch source catalogueswith global photometric calibrationSource re-measurement parameters from consistent list-driven photometry across all available bands in any onefield.

We realise that due to the depth of the VIDEO survey that more precise and survey-specific frame-stackingwill be required. The VIDEO team will undertake this task and provide the ESO archive with the final stackeddata, with merged catalogues from all 5 filters.



The multi-wavelength stacked catalogues will also be made available by the VIDEO team ledby Jarvis and the specific task leaders highlighted in table 6.1, these include VIDEO cataloguesmatched with Spitzer-SWIRE, public Herschel survey data, along with optical, X-ray and radiocatalogues over the VIDEO areas.

6.5 General schedule of the project:T0: Start of observationsT0+12months; Release of science products from first month of survey observationsT0+16month; Release of science products from first 6 months of survey observationsThereafter we would hope that science products can be released within 6 months of raw data arriving in theUK. Optional reprocessing of data based on improved knowledge of instrument would also be considered

7 Envisaged follow-up: (1 page max)

Spectroscopic and small-area deep imaging follow-up

Possibly one of the most important steps for these follow-up studies is the acquisition of redshift information.One of the main aims of VIDEO is to probe galaxies and clusters in the traditional ‘redshift desert’ at 1.3 <z < 1.7, i.e. where there are no bright emission lines in the optical waveband. Therefore, it is both crucialand timely that multi-object near-infrared spectrographs are becoming available on the largest telescopes. Inparticular FMOS on Subaru and EMIR on the GTC in the northern hemisphere and KMOS on the VLT inthe southern hemisphere. Moreover, to probe very distant (z > 2) candidate (proto-)clusters, where only a fewobjects appear to be grouped in the VIDEO survey, will require deeper smaller area follow-up observations.HAWK-I will be the ideal instrument to probe further down the cluster luminosity function of putative z ∼ 2cluster candidates. This is in addition to the instruments currently available on the VLTs and other facilities.For instance VIMOS will continue to be the premier instrument for multi-object spectroscopy in the opticalregime. We will also be pursuing moderate depth optical imaging (probably with the VST) where this doesn’talready exist, to enable even better photometric redshifts, this data will be particularly useful for the z < 1populations. We note that even to cover 15 sq.degrees would not require an extensive time allocation (typically30-60 min integrations per filter per square degree).

From Spitzer to Herschel

Spitzer and Herschel are attempting a complementary exploration of galaxy populations as function of epochand environment - but focusing on the bolometric output from obscured star-formation and AGN activity. It isvital for VIDEO to be in the same fields for scientific reasons (e.g. comparison of different populations in samevolume) and for practical reasons (e.g.. multi-wavelength studies of same galaxies, using VIDEO photometricredshifts for Herschel sources etc.)

The ELAIS-S1, CDF-S and XMM-LSS fields form part of the Spitzer Wide-area InfraRed Extragalactic legacysurvey, (SWIRE, Lonsdale et al. 2003, 2004). The SWIRE survey is now complete (final public data releaseJuly 2006) and covers a total of 49 sq. deg in six fields. These fields were carefully chosen to be in regions oflow galactic dust column density and on this figure of merit they are some of the best fields on the sky. Inthe current plans the three southern SWIRE fields will also be covered in Herschel guaranteed time to a depthof 60mJy at 250, 350 and 500 micron, these observations will pick out some of the most extreme obscuredobjects perfect for study with VISTA and ALMA. In addition one field, probably CDFS or XMM-LSS will becompletely covered to a much greater depth. The data will have a limited proprietary period and the projectteams will be obliged to produce usable data products. The period will be 1 year for observations in the firstyear and six months after that.

(Sub)-millimetre follow-up

(Sub)millimetre observations of key targets detected by VIDEO will constrain the dust and molecular gas con-tent of objects in the very early universe. Observations using (sub)millimetre bolometers (JCMT/SCUBA2,APEX/LABOCA and IRAM/MAMBO), in addition to the Spitzer/Herschel observations, are critical to con-



strain the total FIR luminosities of high-redshift sources. It is of particular importance to derive the propertiesof the molecular gas, from which stars form, in systems in the very early universe (i.e. gas mass, dynamics andexcitation). All the fields targeted by VIDEO will also be reachable with ALMA, making sensitive and highresolution follow-up CO observations feasible, e.g. for the highest-redshift (z > 6) QSOs observations of themolecular gas phase provide invaluable clues regarding the masses of the host galaxies as well as the reservoirof gas that may ultimately form stars. These observations are also of fundamental importance in deriving thedynamical masses of these objects and thus help to put limits on deviations of the M-σ relation as a functionof z and constrain ΛCDM models/simulations in the early stages of galaxy assembly.

Galaxy clusters in the X-ray and the SZ effect

As discussed in section 2.1, there is a clear synergy between VIDEO and X-ray/SZE cluster surveys. Our surveyregions have been chosen to leverage existing X-ray cluster surveys, e.g. XMM-LSS, and also to allow sufficientRA and Dec coverage to be accessible by the upcoming SZE optimised instruments. The existence of VIDEOdata will precipitate requests by this team, and the rest of the X-ray community, for new X-ray surveys in thoseregions not already well covered by XMM or Chandra. Looking to the future, VIDEO will produce exactly thesort of high redshift clusters, AGN and quasars that will be needed as targets for the next generation of X-raysatellites (e.g. XEUS, Con-X). The SZE is a particularly powerful method to discover high redshift clusters, butit requires exhaustive optical/SZ follow-up to secure identifications and redshifts. It is only natural that SZEteams will want to survey the VIDEO regions, in order to get such follow-up “for free”. VIDEO plus SZE willbe a uniquely powerful method to select clusters at the very epoch of formation.

Extension of VIDEO

It is clear that to reap the full benefits of the multi-wavelength coverage over the whole of the SWIRE fieldsdeep near-infrared imaging will be crucial. Thus, we envisage that VIDEO could continue beyond the nominal5 years proposed here as VISTA will still be the premier instrument for conducting this survey, as space-basedprojects such as WISE will not have the spatial resolution to determine galaxy morphologies nor have theshorter near-infrared wavelength coverage.

8 Other remarks, if any: (1 page max)

References

Adelman-McCarthy J.K., et al., 2006, ApJS, 162, 38Albrecht A., et al., 2006 (astro-ph/0609591)Babbedge T.S.R., et al., 2006, MNRAS, submittedBaugh C.M., Lacey C.G., Frenk C.S., Granato G.L., Silva L., Bressan A., Benson A.J., Cole S., 2005, MNRAS,356, 1191Becker R.H., et al., 2001, AJ, 122, 2850Caputi K.I., McLure R.J., Dunlop J.S., Cirasuolo M., Schael A.M., 2006, MNRAS, 366, 609 Caputi K.I., DunlopJ.S., McLure R.J., Roche N.D., 2005, MNRAS, 361, 607Carlstrom J.E., Holder G.P., Reese E.D., 2002, ARA&A, 40, 643Cimatti A., Daddi E., Renzini A., 2006, A&A, 453, 29Colless M., et al., 2001, MNRAS, 328, 1039de Propris R., Stanford S.A., Eisenhardt P.R., Dickinson M., Elston R., 1999, AJ, 118, 719Drory N., Bender R., Feulner G., Hopp U., Maraston C., Snigula J., Hill G.J., 2004, ApJ, 608, 742Hambly N.C. et al., 2004, Proc. SPIE, vol. 5493, 423Emerson J.P. et al., 2004, Proc. SPIE, vol. 5493, 401Evrard A.E., et al., 2002, ApJ, 573, 7Fan X., et al., 2001, AJ, 122, 2833Franx M., et al., 2003, ApJ, 587, 79Hewett P.C., Warren S.J., Leggett S.K., Hodgkin S.T., 2006, MNRAS, in press (astro-ph/0601592)Ilbert O., et al., 2006, A&A submitted, (astro-ph/0603217)



Jarvis M.J., Rawlings S., Eales S.A., Blundell K.M., Bunker A.J., Croft S., McLure R.J., Willott C.J., 2001,MNRAS, 326, 1585Kochanek C.S., et al., 2001, ApJ, 560, 566Lidman C., Rosati P., Demarco R., Nonino M., Mainieri V., Stanford S.A., Toft S., 2004, A&A, 416, 829Lonsdale C., et al., 2003, PASP, 115, 897Lonsdale C., et al., 2004, ApJS, 154, 54Madau P., et al., 1996, MNRAS, 283, 1388McLure et al., 2006, MNRAS, in press (astro-ph/0606116)Martin C.D., et al., 2005, ApJ, 619, 59Martinez-Sansigre A., et al., 2005, Nature, 436, 666Martinez-Sansigre A., et al., 2006, MNRAS, 370, 1479Mo H.J., & White S.D.M., 2002, MNRAS, 336, 112Mullis C.R., Rosati P., Lamer G., Bohringer H., Schwope A., Schuecker P., Fassbender R., 2005, ApJ, 623, 85Richards G.T., et al., 2006, AJ submitted, (astro-ph/0601434)Rosati P., Stanford S.A., Eisenhardt P.R., Elston R., Spirad H., Stern D., Dey A., 1999, AJ, 118, 76Rosati P., Borgani S., Norman C., 2002, ARA&A, 40, 539Saunders W., et al., 2000, MNRAS, 317, 55Springel V., et al., 2005, Nature, 425, 629Somerville R., et al., 2004, ApJ, 600, 171Stanford S.A., et al., AJ, 123, 619Stanford et al., 2005, ApJ, 634, 129Steidel C.C., Giavalisco M., Pettini M., Dickinson M., Adelberger K.L., 1996, ApJ, 462, 17Steidel C.C., Adelberger K.L., Giavalisco M., Dickinson M., Pettini M., 1999, ApJ, 519, 1Thomas D., Maraston C., Bender R., de Oliveira C.M., 2005, ApJ, 621, 673Ueda Y., Akiyama M., Ohta K., Miyaji T., 2003, ApJ, 598, 886van Breukelen C., et al., 2006, MNRAS, (astro-ph/0608624)Willott C.J., Rawlings S., Jarvis M.J., Blundell K.M., 2003, MNRAS, 339, 173


1 vista deep extragalactic observations (video) surveyvideo/public/elais-s1_files/... ·...

Documents